Sourcing

When to Split One Promo Order Across Two Factories

One PO only works when the supplier truly controls the production risk

A single supplier is easier to manage: one PO, one deposit, one artwork file trail, one shipment plan, and usually one pickup. That simplicity only saves money when the factory actually owns the critical processes for the SKUs you are buying. If a supplier is strong in stamped or die-cast metal but brokers patches, lanyards, or kitting to outside workshops, the apparent convenience often turns into extra sample loops, wider quality variation, delayed pack-out, and less visibility when something fails.

For a common promo assortment such as 2,500 soft enamel pins, 800 zinc-alloy keychains, 1,000 woven patches, 5,000 sublimation lanyards, and 500 magnets, outsourced categories typically add 3-7 calendar days per family through artwork forwarding, pre-production sample approval, inter-factory transfers, and final repacking. If the event date is fixed, that delay matters more than a nominal unit-price reduction. Air-freighting even one late SKU can cost USD 450-1,400 for 4-6 cartons on common Asia-US lanes, and a partial reship can erase the savings from putting everything under one quote.

The decision is not one factory versus two. It is whether your supplier has real capability overlap across the order. A qualified metal factory may be excellent for 35 mm stamped soft enamel pins at 1.5 mm thickness, 50 mm die-struck coins at 3.0 mm, and 55 mm die-cast keychains with split rings and jump rings, yet still be a weak controller of woven patches or jacquard lanyards if those are subcontracted. Compare suppliers on delivered performance: in-house process control, MOQ efficiency by SKU, inspection depth, packaging coordination, and total production-to-ship timeline.

Group SKUs by process family, not product label

Start by mapping each item to its actual manufacturing process. That determines tooling method, dimensional tolerance, color control, likely defect modes, and lead time. Product names are too broad to make sourcing decisions. Soft enamel pins, hard enamel badges, coins, cast keychains, and stamped metal magnets are related metal-process items. Woven patches, embroidered patches, jacquard lanyards, screen-printed lanyards, PVC wristbands, and sewn labels are not, even if they all sit inside one campaign kit.

Keep items together when they share material behavior, finishing logic, and inspection criteria. Stamped brass pins, iron challenge coins, and zinc-alloy keychains can usually stay with one metal supplier if die making, stamping or casting, grinding, polishing, plating, color fill, baking, and final assembly are controlled in-house. For 2026 planning, repeat metal runs commonly ship in 10-15 production days after sample approval; new dies or molds usually add 3-6 days. Typical working tolerances are plus or minus 0.10 mm on stamped thickness, plus or minus 0.15 mm on cast thickness, and plus or minus 0.50-1.00 mm on finished length or width depending on geometry.

Split the order when a supplier has to bridge into a different process family. Woven patches require yarn denier selection, stitch-density control, edge finishing by merrow or laser cut, and Pantone-to-thread conversion. Lanyards require webbing or satin selection, print registration, attachment hardware, breakaway placement, and basic tensile checks. Those QC points are unrelated to plating adhesion, enamel fill coverage, edge polishing, pin-post solder strength, or jump-ring closure on metal goods.

Process family	Best sourcing logic	Typical 2026 benchmarks
Stamped pins, coins, metal keychains	Keep with one metal factory if tooling, plating, enamel, and assembly are in-house	MOQ 100-300 for coins, 300-500 for pins, 500-1,000 for keychains; 10-18 production days; FOB USD 0.36-0.82 for 3,000 pcs 35 mm soft enamel pins depending on plating, backing, and epoxy
Metal magnets plus badges	Usually keep together	Shared die-line and finishing control; magnet attachment adds 1-2 days; specify pull force such as 650-900 gf if functional use matters
Woven and embroidered patches	Keep with one textile specialist	MOQ 100-300 workable, 500-1,000 more economical; 9-16 production days; size tolerance usually plus or minus 1.0-2.0 mm depending on border type
Printed or jacquard lanyards and wristbands	Keep with one strap or textile supplier	MOQ 500-1,000; 8-15 production days; standard widths 15 mm and 20 mm; FOB USD 0.22-0.54 for 5,000 pcs sublimation lanyards with standard swivel hook
Metal items plus lanyards or patches	Often split	Different machines, different pre-production samples, and higher subcontracting exposure even when quoted together
Low-spec giveaway sets	Single-source can work	Reasonable when artwork is simple, AQL can be looser, and one late SKU will not break the program

Price the subcontracting penalty instead of trusting the bundle quote

Mixed promo orders usually fail where outsourced lines are presented as if they were factory-owned lines. If a supplier owns die-striking presses, polishing lines, plating tanks, and enamel stations but buys patches and lanyards from partner factories, you should treat the difference as a subcontracting penalty. It may not appear on the quote, but it appears in the schedule, approval burden, and defect risk.

Use three planning assumptions for every outsourced category. First, add 3-7 calendar days to likely lead time for artwork transfer, sample review, inter-factory movement, and repack. Second, add one more approval loop, especially where Pantone conversion, woven detail, or dye-sublimation registration matter. Third, assume weaker root-cause visibility if defects appear, because the contracting supplier may only inspect finished goods rather than control machine settings, incoming materials, or in-line checkpoints.

A realistic quote set makes this visible. A metal factory may quote 3,000 custom 35 mm soft enamel pins at FOB USD 0.42-0.76 each and 1,000 die-cast keychains at FOB USD 0.78-1.32 each; those numbers can be competitive if the production is genuinely in-house. The same supplier may also quote 1,000 woven patches at FOB USD 0.31-0.60 and 5,000 sublimation lanyards at FOB USD 0.24-0.52, but if those lines are brokered, the true cost includes slower sample turnaround, higher carton-mix risk, barcode errors, and a greater chance that final kit assembly misses the booking cutoff.

That is why splitting one category to a specialist often lowers total landed cost even when unit price rises by 3-8 percent. A stronger patch supplier that holds plus or minus 1.0 mm size tolerance, matches thread colors more accurately, and ships on day 12 instead of day 18 can save more than the markup by avoiding partial air freight, repacking fees, chargebacks, or remake claims.

Use MOQ tiers, tooling cost, and calendar days to make the split call

MOQ mismatch is one of the clearest reasons to split. Different product families become economical at different volumes because tooling, setup, labor content, and finishing steps are different. Buyers often overbuy the weak category just to make a bundled quote look efficient, then carry excess inventory for months.

Take a practical mix: 300 challenge coins for VIP gifting, 2,500 enamel pins for channel partners, 800 woven patches for merch packs, and 5,000 event lanyards. Coins remain viable with a metal specialist because they still price reasonably at 100-300 pieces. Lanyards usually sharpen after 1,000 units and often quote best at 3,000-5,000 because setup cost is spread over more units. Woven patches often sit in the middle: possible at 300, clearly more efficient at 500-1,000, and usually not best placed with a factory that mainly runs metal.

Useful 2026 working ranges: soft enamel pins become meaningfully more efficient from 1,000 units upward, though 300-500 is common; die-cast keychains usually improve around 500-1,000 units; challenge coins work well at 100-300 units; woven patches sharpen after 500-1,000 units; sublimation lanyards usually quote best above 1,000 units. New tooling commonly adds USD 45-120 per design for simple stamped metal, USD 80-180 for die-cast shapes, USD 25-70 for woven patch setup depending on edge and backing, and USD 20-60 for standard lanyard setup depending on width, hardware, and individual bagging.

Lead time follows those economics. A repeat enamel pin run is often 10-14 production days; a new mold keychain commonly needs 14-20 days; woven patches usually run 10-16 days; sublimation lanyards often run 8-12 days for 15 mm or 20 mm widths with standard hook and safety buckle. If a fast textile item is trapped inside a slower molded-metal schedule, keeping everything with one vendor can slow the whole order to the pace of the longest path.

Give the tightest-spec SKUs to specialists, not to generalists

Do not split by internal convenience or by whichever item looks most visible in the presentation. Split by technical sensitivity and failure cost. The most expensive problems usually come from items with fine detail, moving parts, exact color requirements, or functional performance, because these trigger remakes, returns, or brand damage faster than simple giveaways.

High-sensitivity metal examples include hard enamel or imitation-hard-enamel pins with line widths below 0.25-0.30 mm, text below about 1.2 mm cap height, polished plated borders that must stay clean after buffing, and dual-post badges that must resist rotation. Functional magnets may require a defined pull force such as 650-900 gf on a specified steel panel. Keychains with lobster clasps, hinged pieces, spinner centers, bottle-opener features, or wire rings also need tighter control on wire diameter, closure gap, rivet fit, and movement smoothness.

Those SKUs deserve tighter QC language. For a sensitivity-critical line, use a pre-production sample plus final inspection at AQL 1.0 or 1.5 for major defects and AQL 2.5 for minor defects. For lower-risk filler lines such as a simple one-color conference lanyard used for a two-day event, AQL 2.5 major and 4.0 minor is often commercially workable if the artwork is simple and the use window is short.

• Split any SKU with line widths under 0.30 mm, text under 1.2 mm cap height, moving parts, or defined pull-force requirements if the supplier cannot show recent comparable samples
• Keep metal SKUs together when one factory controls tooling, plating, enamel fill, curing, and final assembly directly
• Treat retail-ready kitting as a separate capability; a good manufacturer is not automatically a good pack-out operator
• Avoid splitting low-risk filler items when extra cartons, relabeling, consolidation, and receiving fees exceed the quality benefit
• Ask for numeric tolerances on size, thickness, plating, attachment strength, and finish before awarding any quality-sensitive line

Tolerance language should be numeric, not generic. For small metal items, buyers commonly specify plus or minus 0.10 mm thickness on stamped parts and plus or minus 0.15 mm on cast parts. Decorative plating is often specified around 0.03-0.08 micron depending on finish and process; premium wear-facing finishes may target higher values. Patch size tolerance is commonly plus or minus 1.0-2.0 mm depending on border style and shape. Without numbers, two suppliers can both claim compliance while shipping visibly different results.

Build the logistics plan from final pack-out backward

The sourcing model is only as strong as the packaging plan. Two factories can each ship acceptable goods and still create downstream cost if inner-pack counts, barcode placement, carton markings, or ship dates do not align. Warehouses and 3PLs increasingly charge for split receipts, relabeling, and hand repacks, so these costs need to sit inside the sourcing decision from the start.

Work backward from the final deliverable. If the program ships as a campaign kit, define whether each item needs a polybag thickness such as 1.5-2.0 mil, a backing card size, suffocation warning text, UPC or EAN placement, country-of-origin marking, and master-carton quantity. Set carton limits up front, for example keeping export cartons below 15-18 kg gross weight and within agreed dimension caps to reduce crush risk and simplify receiving. For pins and magnets, confirm whether items are mounted to cards before bagging or bagged loose; for lanyards, confirm fold style, banding, and hook orientation.

Transit timing matters as much as production timing. If one supplier finishes on day 11 and the other on day 21, early goods either wait in storage, ship separately, or are repacked twice. For ocean freight, a 5-7 day completion gap may be acceptable. For a trade show, retail reset, or distributor launch with a fixed in-hands date, the same gap can wipe out the gain from splitting. Compare models using delivered timeline: production days plus consolidation, export documents, booking cutoff, customs risk, and freight mode, not just factory completion date.

Decision factor	One factory usually wins when	Split factories usually win when
Process fit	Most SKUs share metal finishing or share textile processes	Order spans metal, textile, print, and function-heavy specialty items
Subcontracting exposure	Supplier owns main equipment and controls final QC directly	Two or more key SKUs are outsourced or only lightly supervised
MOQ efficiency	Volumes align with one supplier's efficient tiers	At least one SKU sits below that supplier's economic floor or tooling sweet spot
Quality sensitivity	Artwork is simple and tolerances are forgiving	Fine detail, moving parts, pull-force, or finish consistency are critical
Pack-out complexity	Bulk packing or simple carding is acceptable	Retail assortments, exact barcode placement, or coordinated kit assembly are required
Schedule risk	Launch window has 5-7 days of slack after production	One late SKU breaks a fixed event, retail reset, or distributor delivery date

Run a split model with one owner, one control sheet, and a hard decision rule

A split-factory structure only works when one person on the buyer side owns the full specification and schedule. That owner controls artwork lock, pre-production approvals, packaging specs, inspection levels, ship window, and exception handling across every supplier. Without that role, each factory optimizes its own workflow and the buyer absorbs the mismatch.

Use one control sheet across all SKUs. At minimum it should list item size, thickness, base material, finish, Pantone references, hardware, packaging type, barcode requirement, master-carton quantity, and inspection level. For metal items, specify brass, iron, or zinc alloy; plating finish; whether epoxy is required; attachment type such as butterfly clutch, rubber clutch, magnet, or safety pin; and tolerance targets. For textile items, specify width, weave or print method, edge finish, hardware, fold style, and individual bagging spec. Also document sample lead time, mass-production lead time in calendar days, and whether final inspection is buyer-arranged, factory-arranged, or third-party.

Commercial ownership matters too. Decide in writing who pays for consolidation, who carries liability if one SKU misses the ship date, whether invoices are paid separately or through a lead vendor, and who can approve substitutions if hardware or packaging runs short. Most mixed-order disputes are not purely about quality. They are accountability disputes created by vague ownership of schedule, consolidation, and final acceptance.

A practical decision rule is straightforward. Keep one supplier when at least four conditions are true: the SKUs are in the same process family, the supplier owns the core equipment, MOQ tiers are aligned, packaging is simple, and the ship date has at least a 7-day buffer after production. Split the order when two or more SKUs are clearly outsourced, one line has narrow visual or functional tolerances, MOQ tiers are mismatched, or the timeline is unforgiving. In most cases, the best structure is not three or four suppliers. It is one metal specialist plus one textile specialist, with written pack-out instructions and a clear consolidation plan.

Have a project? Send your artwork and target quantity and we’ll reply with a detailed quotation within 12 working hours.